Gnome-session (and other gnome programs) appear to be unable to lookup session information. In particular, .gnomerc-errors accumulates messages suggesting in one way or another that it was impossible to get the "session information". Specific and most glaring error is: gnome-session[13185]: WARNING: Unable to determine session: Unable to lookup session information for process '13185' Reproducible: Always Steps to Reproduce: 1. Install gnome-session-2.6.2 2. Login and run programs (terminals, firefox, etc.) as normal. 3. Note accumulation of errors in .gnomerc-errors as well as "strange" behavior in various programs from time-to-time. Actual Results: Errors accumulate in .gnomerc-errors involving the failure of various programs to identify or connect to the current gnome session. Expected Results: Errors should not accumulate. In particular, the program "gnome-session" should *not* have difficulty looking up information about itself. I have previously reported the problem of "losing workspace names" which still exists on my system though not for all users (so its a strange dbus/consolekit/gconf problem). Currently I'm having a problem that gnome-screensaver will not accept a valid password to reactivate the screen and has to be killed from a different VT to get the gnome-session back. This appears to be a dbus error in that is generated by "gnome-session/gsm-consolekit.c" in the function get_current_session_id() via a "dbus_connection_send_with_reply_and_block()" call. I still have not figured out where the "lookup session information" part of the error comes from but I suspect it is from dbus-daemon. In doing some searches, this isn't particularly a Gentoo bug as error reports of a similar nature seem to be showing up in Debian, Ubuntu, ArchLinux, etc. all over the last year or so seeming to involve Gnome upgrades. There is a hint that this may be a "policykit" problem (my gnome-session is built "-policykit"). Looking back through the ebuild's it looks like in gnome 2.22 there was no concept of "policykit", it appeared in 2.24 and it now may be more or less mandatory in 2.26. (Though it does not appear that the portage keywords/use files make it mandatory if one tries to work with 2.26.) There would be 2 bugs here: 1) Gnome 2.26 programs should all work properly without policykit (since it adds at least 1 MB to a gnome install and since I'm running a single user system and have no real need for "policies"). 2) Gentoo should adjust the ebuild upgrading process (typical USE options) so that when a user has a complete "bleeding edge" install it works as well as the previous install (e.g. 2.24 or 2.22). In theory that may mean that things like hal, dbus, policykit and lord knows what else go from being "optional" to "mandatory". It would be nice if one or more of the Gentoo gnome developers/users would post an "ideal" cutting edge configuration (i.e. all the use flags for all gnome/common utilities) that enable running a personal desktop system for a week or more and getting no errors in .gnomerc-errors.
It's a warning, ignore it. Or use policykit.
Please report only one issue per bug, your reports are seriously over complicated and hinder our capacity to actually solve your problems. It is most likely that this bug is yet another duplicate of your other reports (bug #280329 and bug #277361) in which we mostly see that you are doing weird things with your system that we are unlikely to figure out given the shallow informations you provide. I'm not saying we won't help you fix it, but please focus on one thing at a time and provide us with the informations we asked in the other bugs.
(In reply to comment #2) > I'm not saying we won't help you fix it, but please focus on one thing at a > time and provide us with the informations we asked in the other bugs. > Also, please don't be so verbose. The more you write, the more difficult it is for us to figure out what you're trying to say. Be brief and succinct. If we need more test cases/information, we'll ask you. More words != more information. However, more words to read == more time consumed.
